The purpose of this proposal is to develop statistical methodology which is particularly appropriate for the study of the epidemiology of AIDS and for clinical research in AIDS, with an emphasis towards markers of disease progression. Four distinct projects are proposed, these are (i) developing semi-parametric methods for use in the mother-to-child (HIV transmission problem, (ii) developing statistical methods for analyzing longitudinal data or performing survival analysis in the presence of missing covariates, (iii) to investigate non-parametric schemes for estimating normal ranges from repeated measurements data, and (iv) developing statistical methods for estimating the distribution of a marker value, which is measured with error, at the time of an endpoint.